Arch Jignesh Desai LinkedIn Github Portfolio
Boston, MA adir4b@r.postjobfree.com 979-***-****
EDUCATION
Master’s in Industrial Engineering (Data Science Specialization) Dec 2020 Texas A&M University GPA: 4.00/4.00 College Station, Texas Bachelor’s in Mechanical Engineering May 2017
Sardar Vallabhbhai National Institute of Technology (SVNIT) GPA: 8.76/10.00 Surat, India TECHNICAL SKILLS
Skills: Machine Learning, Deep Learning, Time Series Analysis, NLP, Statistics, A/B Testing, Big Data, Explainable AI Languages: (Pandas, Numpy, Scikit-Learn, Scipy, Keras, Matplotlib), R (Dplyr, Tidyr, Caret, Ggplot2),, C++ Tools: MySQL,, Git, PySpark, Hadoop, MapReduce, Amazon Web Services (AWS), Flask, MS Excel EXPERIENCE
National Renewable Energy Laboratory Oct 2019 - Aug 2020 Data Science Intern Denver, Colorado
Extracted 10+ features using wind turbine gearbox design and physics-based models that reduced false alarms by 50% and improved F1 score by 12%; Published a paper for the same as a first author at PHM Society [Link].
Established an ETL pipeline to manipulate data using PySpark,, and AWS; Built machine learning models for gearbox bearing failure prognosis using Logistic Regression, Random Forest, XGBoost, 1D CNN, and LSTM.
Implemented robust anomaly detection systems using regression model residuals, eigenvalues, and Autoencoders.
Carried out ad-hoc data analysis, statistical tests, and fault analysis; Delivered insights using dashboards. Dassault Systèmes Jul 2017 - Jul 2018
R&D Associate Engineer Pune, India
Conducted performance analysis of Feature APIs and proposed solutions which can reduce run-time by 7%.
Revamped and improved front-end SolidWorks Pattern APIs for easier user access and better interpretability.
Developed, optimized, and debugged SolidWorks APIs for better CAD model implementations using C++, C#, and VBA. Indian Institute of Technology May 2016 - Jul 2016 Research Assistant Delhi, India
Executed numerical analysis of asymmetrically heated tilted channels using simulated data from ANSYS Fluent.
Identified the best channel attributes for maximum heat transfer using statistical tests and data analysis in R. PROJECTS
Customer Survival Analysis and Churn Prediction (Lifelines, Scikit-Learn, Matplotlib, Flask, SHAP, Eli5) [App] [Link]
Employed Kaplan-Meier estimator and Cox Proportional Hazard model to analyze customer survival over time.
Implemented Random Forest model that predicts customer churn with 0.85 AUC; Built a Flask App which shows churn probabilities, Lifetime Value, cumulative hazard, survival curve, and SHAP values based on customer data. Predictive Maintenance of an Engine (Keras, Scikit-Learn, SHAP, Matplotlib) [Link]
Built exponential degradation model, similarity-based prognostic approach, and LSTM model to predict engine RUL.
Developed LSTM, RNN, 1D CNN, and CNN-SVM models to predict engine failure 50 cycles ahead of its time with 0.92 F1 score; Utilized Sensitivity analysis and SHAP values to find important features from these black-box models. Instacart Market Basket Analysis (Mlxtend, Scikit-Learn, XGBoost, Matplotlib) [Link]
Analyzed 3 million grocery orders and performed segmentation and affinity analysis to study user purchase patterns.
Extracted 30+ features from customer transactional data and built XGBoost model that predicts which previously purchased item will be in user’s next order with 0.83 AUC. News Articles Recommendation (Tweepy, NLTK, Gensim, Newspaper, Matplotlib) [Link]
Scraped and cleaned Twitter data; Grouped users using TF-IDF, cosine similarity, and K-Means clustering algorithm.
Developed feature combination hybrid-filtering recommendation system to suggest news articles to a Twitter user based on LDA Topic Modelling Normalized probabilities, sentiment score, and Jaccard similarity. ACHIEVEMENTS
Winner of a TAMU Datathon 2020 among 50+ teams.
Recipient of TAMU Scholarship and Fee Waiver for excellent academic performance.
Vice President of Events at Indian Graduate Students Association (400+ members).